| Message ID | 1396025007-27592-8-git-send-email-alex.bennee@linaro.org |
|---|---|
| State | Superseded |
| Headers | show |
On 28 March 2014 16:43, Alex Bennée <alex.bennee@linaro.org> wrote: > The current exec and cpu flags only show information when we enter a > TranslationBlock. The blocks will then chain together until there is a > reason to exit which can make things harder to follow. This new option > disables the chaining for debugging purposes. > > + { CPU_LOG_TB_NOCHAIN, "nochain", > + "disable chaining of translation blocks (better tracing, but slower)" }, I know it's a convenient place to put it, but it's not really a tracing option so I'm a bit dubious about putting it in the log flags... thanks -- PMM
Peter Maydell <peter.maydell@linaro.org> writes: > On 28 March 2014 16:43, Alex Bennée <alex.bennee@linaro.org> wrote: >> The current exec and cpu flags only show information when we enter a >> TranslationBlock. The blocks will then chain together until there is a >> reason to exit which can make things harder to follow. This new option >> disables the chaining for debugging purposes. >> >> + { CPU_LOG_TB_NOCHAIN, "nochain", >> + "disable chaining of translation blocks (better tracing, but slower)" }, > > I know it's a convenient place to put it, but it's not really > a tracing option so I'm a bit dubious about putting it in > the log flags... Yeah I know, but I didn't really want to make a new tcg opts set of flags as it's more of logging clarity kinda option. Of course this glosses over the fact it is also useful for tcg trace-events (when I send them). Is it time to expose a set of TCG fiddling variables? > > thanks > -- PMM
diff --git a/cpu-exec.c b/cpu-exec.c index abe02b7..c20349c 100644 --- a/cpu-exec.c +++ b/cpu-exec.c @@ -620,8 +620,10 @@ int cpu_exec(CPUArchState *env) } /* see if we can patch the calling TB. When the TB spans two pages, we cannot safely do a direct - jump. */ - if (next_tb != 0 && tb->page_addr[1] == -1) { + jump. It can also complicate tracing so disable if + asked (but be aware this can change behaviour). */ + if (next_tb != 0 && tb->page_addr[1] == -1 && + !qemu_loglevel_mask (CPU_LOG_TB_NOCHAIN)) { tb_add_jump((TranslationBlock *)(next_tb & ~TB_EXIT_MASK), next_tb & TB_EXIT_MASK, tb); } diff --git a/include/qemu/log.h b/include/qemu/log.h index f6cef9e..82902b0 100644 --- a/include/qemu/log.h +++ b/include/qemu/log.h @@ -40,6 +40,7 @@ static inline bool qemu_log_enabled(void) #define CPU_LOG_RESET (1 << 9) #define LOG_UNIMP (1 << 10) #define LOG_GUEST_ERROR (1 << 11) +#define CPU_LOG_TB_NOCHAIN (1 << 12) /* Returns true if a bit is set in the current loglevel mask */ diff --git a/qemu-log.c b/qemu-log.c index a54b332..dc78a93 100644 --- a/qemu-log.c +++ b/qemu-log.c @@ -175,6 +175,8 @@ const QEMULogItem qemu_log_items[] = { { LOG_GUEST_ERROR, "guest_errors", "log when the guest OS does something invalid (eg accessing a\n" "non-existent register)" }, + { CPU_LOG_TB_NOCHAIN, "nochain", + "disable chaining of translation blocks (better tracing, but slower)" }, { 0, NULL, NULL }, };
The current exec and cpu flags only show information when we enter a TranslationBlock. The blocks will then chain together until there is a reason to exit which can make things harder to follow. This new option disables the chaining for debugging purposes. Signed-off-by: Alex Bennée <alex.bennee@linaro.org>